wecharyu commented on PR #4744: URL: https://github.com/apache/hive/pull/4744#issuecomment-1877318065
Hi @dengzhhu653 , I agree with @zhangbutao that we should also clear the schema in this patch. If we want a smooth upgrade for this feature, we need do as follows: Stage1: new_version1 read stat by join id and still write all redundant names (just as you suggest), old_version read by names, Stage2: new_version2 write without redundant names, both new_version2 and new_version1 read by join id Stage3: remove the redundant columns in schema (new_version3 for schema update) We will need three patches (versions) to achieve this feature, which seems a bit tedious. If we combine all changes in this patch, and upgrade HMS clusters one by one, and finally change the schema. Then the only issue was that the old HMS instances could not read the new stats created by new HMS instances, I think it's acceptable. Furthermore, we can offline all old instances in the process of upgrade to avoid this issue. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
